simonmichael/hledger
1
1
Fork 0
mirror of https://github.com/simonmichael/hledger.git synced 2024-11-08 07:09:28 +03:00
Code Issues Projects Releases Wiki Activity
ec2436025c
hledger/hledger-lib/CHANGES.md
Simon Michael 68e7ff4644 ;doc: finalise changelogs
2021-07-03 14:51:51 -10:00

1344 lines
49 KiB
Markdown
Raw Blame History

<!--
_ _ _
| (_) |__
| | | '_ \
| | | |_) |
|_|_|_.__/
-->
Internal/api/developer-ish changes in the hledger-lib (and hledger) packages.
For user-visible changes, see the hledger package changelog.
# 1.22 2021-07-03
- GHC 9.0 is now officially supported, and GHC 8.0, 8.2, 8.4 are not;
building hledger now requires GHC 8.6 or greater.
- Added now-required lower bound on containers. (#1514)
- Added useColor, colorOption helpers usable in pure code, eg for debug output.
- Added a Show instance for AmountDisplayOpts and WideBuilder, for debug logging.
Many internal refactorings/improvements/optimisations by Stephen Morgan,
including:
- Don't infer a txn price with same-sign amounts. (#1551)
- Clean up valuation functions, and make clear which to use where. (#1560)
- Replace journalSelectingAmountFromOpts with journalApplyValuationFromOpts.
This also has the effect of allowing valuation in more reports, for
example the transactionReport.
- Refactor to eliminate use of printf.
- Remove unused String, Text utility functions.
- Replace concat(Top|Bottom)Padded with textConcat(Top|Bottom)Padded.
- Export Text.Tabular from Text.Tabular.AsciiWide, clean up import lists.
- When matching an account query against a posting, don't try to match
against the same posting twice, in cases when poriginal is Nothing.
- Create mixedAmountApplyValuationAfterSumFromOptsWith for doing any
valuation needed after summing amounts.
- Create journalApplyValuationFromOpts. This does costing and
valuation on a journal, and is meant to replace most direct calls of
costing and valuation. The exception is for reports which require
amounts to be summed before valuation is applied, for example a
historical balance report with --value=end.
- Remove unused (amount|mixedAmount|posting|transaction)ApplyCostValuation functions.
- Remove unnecessary normalisedMixedAmount.
- Remove `showAmounts*B` functions, replacing them entirely with
`showMixedAmount*B` functions.
- Pull "show-costs" option used by the Close command up into ReporOpts.
- Add more efficient toEncoding for custom ToJSON declarations.
- Fix ledgerDateSpan, so that it considers both transaction and
posting dates. (#772)
- Move reportPeriodName to Hledger.Reports.ReportOptions, use it for
HTML and CSV output for compound balance reports.
- Simplify the JSON representation of AmountPrecision. It now uses the
same JSON representation as Maybe Word8. This means that the JSON
serialisation is now broadly compatible with that used before the
commit f6fa76bba7530af3be825445a1097ae42498b1cd, differing only in
how it handles numbers outside Word8 and that it can now produce
null for NaturalPrecision.
- A number of AccountName and Journal functions which are supposed to
produce unique sorted results now use Sets internally to be slightly
more efficient. There is also a new function journalCommodities.
- More efficiently check whether Amounts are or appear to be zero.
Comparing two Quantity (either with == or compare) does a lot of
normalisation (calling roundMax) which is unnecessary if we're
comparing to zero. Do things more directly to save work.
For `reg -f examples/10000x10000x10.journal`, this results in
- A 12% reduction in heap allocations, from 70GB to 62GB
- A 14% reduction in (profiled) time, from 79s to 70s
Results for bal -f examples/10000x10000x10.journal are of the same
order of magnitude.
- In sorting account names, perform lookups on HashSets and HashMaps,
rather than lists. This is probably not an enormous performance sink
in real situations, but it takes a huge amount of time and memory in
our benchmarks (specifically 10000x10000x10.journal). For
`bal -f examples/10000x10000x10.journal`, this results in
- A 23% reduction in heap allocation, from 27GiB to 21GiB
- A 33% reduction in (profiled) time running, from 26.5s to 17.9s
- Minor refactor, using foldMap instead of asum . map . toList.
- Do not call showAmount twice for every posting. For print -f
examples/10000x10000x10.journal, this results in a 7.7% reduction in
heap allocations, from 7.6GB to 7.1GB.
- Some efficiency improvements in register reports.
Use renderRow interface for Register report.
For `reg -f examples/10000x10000x10.journal`, this results in:
- Heap allocations decreasing by 55%, from 68.6GB to 31.2GB
- Resident memory decreasing by 75%, from 254GB to 65GB
- Total (profiled) time decreasing by 55%, from 37s to 20s
- Split showMixedAmountB into showMixedAmountB and showAmountsB, the
former being a simple wrapper around the latter. This removes the
need for the showNormalised option, as showMixedAmountB will always
showNormalised and showAmountsB will never do so.
- Change internal representation of MixedAmount to use a strict Map
instead of a list of Amounts. No longer export Mixed constructor, to
keep API clean. (If you really need it, you can import it directly
from Hledger.Data.Types). We also ensure the JSON representation of
MixedAmount doesn't change: it is stored as a normalised list of
Amounts.
This commit improves performance. Here are some indicative results:
hledger reg -f examples/10000x1000x10.journal
- Maximum residency decreases from 65MB to 60MB (8% decrease)
- Total memory in use decreases from 178MiB to 157MiB (12% decrease)
hledger reg -f examples/10000x10000x10.journal
- Maximum residency decreases from 69MB to 60MB (13% decrease)
- Total memory in use decreases from 198MiB to 153MiB (23% decrease)
hledger bal -f examples/10000x1000x10.journal
- Total heap usage decreases from 6.4GB to 6.0GB (6% decrease)
- Total memory in use decreases from 178MiB to 153MiB (14% decrease)
hledger bal -f examples/10000x10000x10.journal
- Total heap usage decreases from 7.3GB to 6.9GB (5% decrease)
- Total memory in use decreases from 196MiB to 185MiB (5% decrease)
hledger bal -M -f examples/10000x1000x10.journal
- Total heap usage decreases from 16.8GB to 10.6GB (47% decrease)
- Total time decreases from 14.3s to 12.0s (16% decrease)
hledger bal -M -f examples/10000x10000x10.journal
- Total heap usage decreases from 108GB to 48GB (56% decrease)
- Total time decreases from 62s to 41s (33% decrease)
If you never directly use the constructor Mixed or pattern match against
it then you don't need to make any changes. If you do, then do the
following:
- If you really care about the individual Amounts and never normalise
your MixedAmount (for example, just storing `Mixed amts` and then
extracting `amts` as a pattern match, then use should switch to using
[Amount]. This should just involve removing the `Mixed` constructor.
- If you ever call `mixed`, `normaliseMixedAmount`, or do any sort of
amount arithmetic (+), (-), then you should replace the constructor
`Mixed` with the function `mixed`. To extract the list of Amounts, use
the function `amounts`.
- Any remaining calls to `normaliseMixedAmount` can be removed, as that
is now the identity function.
- Create a new API for MixedAmount arithmetic. This should supplant
the old interface, which relied on the Num typeclass. MixedAmount
did not have a very good Num instance. The only functions which were
defined were fromInteger, (+), and negate. Furthermore, it was not
law-abiding, as 0 + a /= a in general. Replacements for used
functions are:
0 -> nullmixedamt / mempty
(+) -> maPlus / (<>)
(-) -> maMinus
negate -> maNegate
sum -> maSum
sumStrict -> maSum
Also creates some new constructors for MixedAmount:
mixedAmount :: Amount -> MixedAmount
maAddAmount :: MixedAmount -> Amount -> MixedAmount
maAddAmounts :: MixedAmount -> [Amount] -> MixedAmount
Add Semigroup and Monoid instances for MixedAmount.
Ideally we would remove the Num instance entirely.
The only change needed have nullmixedamt/mempty substitute for 0
without problems was to not squash prices in
mixedAmount(Looks|Is)Zero. This is correct behaviour in any case.
# 1.21 2021-03-10
- Building Hledger.Data.Journal no longer fails if the monad-extras
package is installed.
- Many parts of the hledger-lib and hledger APIs have become more
Text-ified, expecting or returning Text instead of String, reducing
hledger's time and resident memory requirements by roughly 10%.
Some functions now use WideBuilder (a text "builder" which keeps track
of width), to concatenate text more efficiently. There are some
helpers for converting to and from WideBuilder (wbUnpack, wbToText..)
showAmountB/showMixedAmountB are new amount-displaying functions
taking an AmountDisplayOpts. These will probably replace the old
show(Mixed)Amount* functions. (#1427, Stephen Morgan)
- AtThen valuation is now implemented for all report types.
amountApplyValuation now takes the posting date as an argument.
(transaction/posting)ApplyValuation's valuation type and
transaction/posting arguments have been reordered like
amountApplyValuation's. (Stephen Morgan)
- Amount, AmountPrice, AmountStyle, DigitGroupStyle fields are now
strict. (Stephen Morgan)
- Amount prices are now stored with their sign, so negative prices can
be represented. (They seem to have always worked, but now the
internal representation is more accurate.) (Stephen Morgan)
- normaliseMixedAmount now combines Amounts with TotalPrices in the
same commodity. (Stephen Morgan)
- normaliseMixedAmount now uses a strict Map for combining amounts
internally, closing a big space leak. (Stephen Morgan)
- (multiply|divide)(Mixed)?Amount now also multiply or divide the
TotalPrice if it is present, and the old
(multiply|divide)(Mixed)?AmountAndPrice functions are removed. (Stephen Morgan)
- (amount|mixedAmount)(Looks|Is)Zero functions now check whether both
the quantity and the cost are zero. This is usually what you want,
but if you do only want to check whether the quantity is zero, you
can run mixedAmountStripPrices (or similar) before this. (Stephen Morgan)
- commodityStylesFromAmounts now consumes the list immediately,
reducing the maximum heap size per thread from ~850K to ~430K in a
real-world register report. (Stephen Morgan)
- *ApplyValuation functions take two less arguments, and
*ApplyCostValuation functions have been added, performing both
costing and valuation. (Stephen Morgan)
- traceAtWith now has a level argument and works properly.
- API changes include:
```
Hledger.Data.Amount:
setAmountPrecision -> amountSetPrecision
setFullPrecision -> amountSetFullPrecision
setMixedAmountPrecision -> mixedAmountSetPrecision
showMixed -> showMixedAmountB
showMixedLines -> showMixedAmountLinesB
-mixedAmountSetFullPrecision
Hledger.Data.Journal:
mapJournalTransactions -> journalMapTransactions
mapJournalPostings -> journalMapPostings
-mapTransactionPostings
+journalPayeesUsed
+journalPayeesDeclaredOrUsed
Hledger.Data.Transaction:
+transactionFile
+transactionMapPostings
Hledger.Data.Valuation:
-valuationTypeIsCost
-valuationTypeIsDefaultValue
-ValuationType's AtDefault constructor
Hledger.Query:
+matchesDescription
+matchesPayeeWIP
Hledger.Utils.Text:
+textConcatBottomPadded
+wbToText
+wbUnpack
Text.Tabular.AsciiWide:
alignCell -> textCell
```
# 1.20.4 2021-01-29
- See hledger.
# 1.20.3 2021-01-14
- See hledger.
# 1.20.2 2020-12-28
- Fix the info manuals' node structure.
- Drop unused parsec dependency.
# 1.20.1 2020-12-15
- renamed: updateReportSpecFromOpts -> updateReportSpec[With]
# 1.20 2020-12-05
- added: journalApplyAliases, transactionApplyAliases, postingApplyAliases
- a new more robust price lookup implementation, fgl library dropped (#1402)
- Reverted a stripAnsi change in 1.19.1 that caused a 3x slowdown of amount rendering
in terminal reports. (#1350)
- Amount and table rendering has been improved, so that stripAnsi is no longer needed.
This speeds up amount rendering in the terminal, speeding up some reports by 10% or more since 1.19.
(Stephen Morgan)
- global commodity display styles can now be set in InputOpts or Journal,
overriding all others (declared or inferred). This is used by the import
command and probably command-line options in future.
- Journal keeps a new piece of parsing state, a decimal mark character,
which can optionally be set to force the number format expected by all
amount parsers.
- Remove Empty Query constructor, which does nothing and has done so for a very long time. (Stephen Morgan)
- In ReportOpts, store query terms term-by-term in a list in querystring_. (Stephen Morgan)
This helps deal with tricky quoting issues, as we no longer have to make
sure everything is quoted properly before merging it into a string.
- Implement concat(Top|Bottom)Padded in terms of renderRow, allowing them to be width aware. (Stephen Morgan)
- Expand Tabular.AsciiWide to allow multiline, custom-width,
vertically/horizontally-aligned cells, and optional table borders.
(Stephen Morgan)
- Introduce showMixed*Unnormalised, eliminate most direct calls of strWidth. (Stephen Morgan)
- showMixedAmountElided now makes better use of space, showing as many
Amounts possible as long as they and the elision string fit within
32 characters. (Stephen Morgan)
- Add Functor instance for CompoundPeriodicReport. (Stephen Morgan)
- Generalise CBCSubreportSpec to allow more subreport control. (Stephen Morgan)
- Export some MultiBalanceReport helper functions. (Stephen Morgan)
- Make Default instances clearer, remove Default instance for Bool. (Stephen Morgan)
- Many ReportOpts-related changes, such as the addition of ReportSpec, aimed
at preventing runtime errors (from parsing: regexps, dates, format strings;
from not having today's date set; etc.)
ReportSpec holds a ReportOpts, the day of the report, and the Query generated from these.
- StringFormat now takes an optional overline width, which is
currently only used by defaultBalanceLineFormat. (Stephen Morgan)
- quoteIfNeeded should not escape the backslashes in unicode code points. (Stephen Morgan)
- Export OrdPlus and constructors. (Stephen Morgan)
- Debug output now uses pretty-simple instead pretty-show.
This hopefully gives overall nicer debug output (eg in colour),
including for values which don't have Read-able Show output.
This means that we can start removing custom Show instances
that were a workaround for pretty-show. Eg account names
in debug output no longer show their colons as underscores.
Here's some old pretty-show output:
CsvRules
{ rdirectives = [ ( "skip" , "1" ) ]
, rcsvfieldindexes = [ ( "date" , 1 ) , ( "amount" , 2 ) ]
, rassignments = [ ( "amount" , "%2" ) , ( "date" , "%1" ) ]
, rconditionalblocks = []
}
And the new pretty-simple output:
CsvRules
{ rdirectives=
[ ( "skip", "1" ) ]
, rcsvfieldindexes=
[ ( "date", 1 ), ( "amount", 2 ) ]
, rassignments=
[ ( "amount", "%2" ), ( "date", "%1" ) ]
, rconditionalblocks= []
}
We require pretty-simple 4.0.0.0 to get this compact output.
It's a little less compact than pretty-show, but not too bad.
Non-compact pretty-simple output would be:
CsvRules
{ rdirectives=
[
( "skip"
, "1B"
)
]
, rcsvfieldindexes=
[
( "date"
, 1
)
,
( "amount"
, 2
)
]
, rassignments=
[
( "amount"
, "%2"
)
,
( "date"
, "%1"
)
]
, rconditionalblocks=[]
}
# 1.19.1 2020-09-07
- Allow megaparsec 9
- stripAnsi: correctly strip ansi sequences with no
numbers/semicolons. (Stephen Morgan)
- Added case-insensitive accountNameToAccountRegexCI,
accountNameToAccountOnlyRegexCI, made the default account type
queries case insensitive again. (#1341)
# 1.19 2020-09-01
- Added a missing lower bound for aeson, making cabal installs more
reliable. (#1268)
- The Regex type alias has been replaced by the Regexp ADT, which
contains both the compiled regular expression (so is guaranteed to
be usable at runtime) and the original string (so can be serialised,
printed, compared, etc.) A Regexp also knows whether is it case
sensitive or case insensitive. The Hledger.Utils.Regex API has
changed. (#1312, #1330).
- Typeable and Data instances are no longer derived for hledger's
data types; they were redundant/no longer needed.
- NFData instances are no longer derived for hledger's data types.
This speeds up a full build by roughly 7%. But it means we can't
deep-evaluate hledger values, or time hledger code with Criterion.
https://github.com/simonmichael/hledger/pull/1330#issuecomment-684075129
has some ideas on this.
- Query no longer has a custom Show instance
- Hledger.Utils.String: quoteIfNeeded now actually escapes quotes in
strings. escapeQuotes was dropped. (Stephen Morgan)
- Hledger.Utils.Tree: dropped some old utilities
- Some fromIntegral calls have been replaced with safer code, removing
some potential for integer wrapping bugs (#1325, #1326)
- Parsing numbers with more than 255 decimal places now gives an error
instead of silently misparsing (#1326)
- Digit groups are now limited to at most 255 digits each. (#1326)
- Exponents are parsed as Integer rather than Int.
This means exponents greater than 9223372036854775807 or less than
-9223372036854775808 are now parsed correctly, in theory. (In
practice, very large exponents will cause hledger to eat all your
memory, so avoid them for now.) (#1326)
- AmountStyle's asprecision is now a sum type with Word8, instead of
an Int with magic values.
- DigitGroupStyle uses Word8 instead of Int.
- Partial helper function parsedate has been dropped, use fromGregorian instead.
- Partial helper function mkdatespan has been dropped.
- Helper function transaction now takes a Day instead of a date string. (Stephen Morgan)
- Old CPP directives made redundant by version bounds have been
removed. (Stephen Morgan)
- Smart dates are now represented by the SmartDate type, and are
always well formed. (Stephen Morgan)
- accountTransactionsReport (used for hledger aregister and
hledger-ui/hledger-web registers) now filters transactions more
thoroughly, so eg transactions dated outside the report period will
not be shown. Previously the transaction would be shown if it had
any posting dated inside the report period. Possibly some other
filter criteria now get applied that didn't before. I think on
balance this will give slightly preferable results.
- The old BalanceReport code has been dropped at last, replaced by
MultiBalanceReport so that all balance reports now use the same
code. (Stephen Morgan, #1256).
- The large multiBalanceReport function has been split up and refactored
extensively.
- Tabular data formerly represented as [[MixedAmount]] is now HashMap
AccountName (Map DateSpan Account). Reports with many columns are now faster.
- Calculating starting balances no longer calls the whole balanceReport,
just the first few functions.
- displayedAccounts is completely rewritten. Perhaps one subtle thing to
note is that in tree mode it no longer excludes nodes with zero inclusive
balance unless they also have zero exclusive balance.
- Simon's note: "I'll mark the passing of the old multiBalanceReport, into
which I poured many an hour :). It is in a way the heart (brain ?) of
hledger - the key feature of ledgerlikes (balance report) and a key
improvement introduced by hledger (tabular multiperiod balance reports)
...
Thanks @Xitian9, great work."
# 1.18.1 2020-06-21
- fix some doc typos (Martin Michlmayr)
# 1.18 2020-06-07
- added: getHledgerCliOpts', takes an explicit argument list
- added: toJsonText
- changed: isNegativeMixedAmount now gives an answer for
multi-commodity amounts which are all negative
- changed: multiBalanceReport now gets the query from ReportOpts (Dmitry Astapov)
- renamed:
isZeroAmount -> amountLooksZero
isReallyZeroAmount -> amountIsZero
isZeroMixedAmount -> mixedAmountLooksZero
isReallyZeroMixedAmount -> mixedAmountIsZero
isReallyZeroMixedAmountCost dropped
- renamed: finaliseJournal -> journalFinalise
- renamed: fixedlotpricep -> lotpricep, now also parses non-fixed lot prices
- dropped: transactionPostingBalances
- dropped: outputflags no longer exported by Hledger.Cli.CliOptions
- fixed: documentation for journalExpenseAccountQuery (Pavan Rikhi)
# 1.17.1 2020-03-19
- require newer Decimal, math-functions libs to ensure consistent
rounding behaviour, even when built with old GHCs/snapshots.
hledger uses banker's rounding (rounds to nearest even number, eg
0.5 displayed with zero decimal places is "0").
- added: debug helpers traceAt, traceAtWith
- Journal is now a Semigroup, not a Monoid (since <> is right-biased). (Stephen Morgan)
# 1.17.0.1 2020-03-01
- fix org heading comments and doctest setup comment that were
breaking haddock (and in some cases, installation)
# 1.17 2020-03-01
- Reader-finding utilities have moved from Hledger.Read to
Hledger.Read.JournalReader so the include directive can use them.
- Reader changes:
- rExperimental flag removed
- old rParser renamed to rReadFn
- new rParser field provides the actual parser.
This seems to require making Reader a higher-kinded type, unfortunately.
- Hledger.Tabular.AsciiWide now renders smoother outer borders in
pretty (unicode) mode.
Also, a fix for table edges always using single-width intersections
and support for double horizontal lines with single vertical lines. (Eric Mertens)
- Hledger.Utils.Parse: restofline can go to eof also
- Hledger.Read cleanup
- Hledger.Read.CsvReader cleanup
Exports added: CsvRecord, CsvValue, csvFileFor.
Exports removed: expandIncludes, parseAndValidateCsvRules, transactionFromCsvRecord
- more cleanup of amount canonicalisation helpers (#1187)
Stop exporting journalAmounts, overJournalAmounts, traverseJournalAmounts.
Rename journalAmounts helper to journalStyleInfluencingAmounts.
- export mapMixedAmount
- Don't store leaf name in PeriodReport. (Stephen Morgan)
Calculate at the point of consumption instead.
- Generalise PeriodicReport to be polymorphic in the account labels. (Stephen Morgan)
- Use records instead of tuples in PeriodicReport. (Stephen Morgan)
- Use PeriodicReport in place of MultiBalanceReport. (Stephen Morgan)
- Calculate MultiReportBalance columns more efficiently. (Stephen Morgan)
Only calculate posting date once for each posting, and calculate their
columns instead of checking each DateSpan separately.
- Moved JSON instances from hledger-web to hledger-lib (Hledger.Data.Json),
and added ToJSON instances for all (?) remaining data types, up to Ledger.
- Dropped nullassertion's "assertion" alias, fixing a warning.
Perhaps we'll stick with the null* naming convention.
# 1.16.2 2020-01-14
- add support for megaparsec 8 (#1175)
# 1.16.1 2019-12-03
- Drop unnecessary mtl-compat dependency
- Fix building with GHC 8.0, 8.2
# 1.16 2019-12-01
- drop support for GHC 7.10, due to MonadFail hassles in JournalReader.hs
- add support for GHC 8.8, base-compat 0.11 (#1090)
We are now using the new fail from the MonadFail class, which we
always import qualified as Fail.fail, from base-compat-batteries
Control.Monad.Fail.Compat to work with old GHC versions. If old fail
is needed (shouldn't be) it should be imported qualified as
Prelude.Fail, using imports such as:
import Prelude hiding (fail)
import qualified Prelude (fail)
import Control.Monad.State.Strict hiding (fail)
import "base-compat-batteries" Prelude.Compat hiding (fail)
import qualified "base-compat-batteries" Control.Monad.Fail.Compat as Fail
- hledger and hledger-lib unit tests have been ported to tasty.
- The doctest suite has been disabled for now since it doesn't run
well with cabal (#1139)
# 1.15.2 2019-09-05
Changes:
- postingApplyValuation, mixedAmountApplyValuation, amountApplyValuation
take an argument, the report end date if one was specified.
# 1.15.1 2019-09-02
- fix failing doctests
# 1.15 2019-09-01
Removals include:
- journalPrices
- BalanceHistoryReport
- postingValueAtDate
Additions include:
- MarketPrice (more pure form of PriceDirective without the amount style information)
- PriceOracle (efficient lookup of exchange rates)
- ValuationType (ways to convert amount value)
- aliasnamep (export)
- setNaturalPrecisionUpTo
- dbgNWith, ptraceAtWith
- postingTransformAmount, postingToCost, postingValue
- amountToCost, mixedAmountToCost
- valueTypeFromOpts
- mapJournalTransactions, mapJournalPostings, mapTransactionPostings
- journalStartDate, journalEndDate
- journalPriceOracle
- marketPriceReverse
- priceDirectiveToMarketPrice
- mixedAmountApplyValuation
- mixedAmountValueAtDate
Changes include:
- Price -> AmountPrice, AKA "transaction price"
- old MarketPrice -> PriceDirective
- TransactionsReport/AccountTransactionsReport split into separate files
- journalTransactionsReport -> transactionsReport
- accountTransactionsReportItems: rewrite using catMaybes and mapAccumL (Henning Thielemann)
- optionally save the current date in ReportOpts
- Hledger.Cli tests now have correct prefix; add Cli.Utils tests
- MultiBalanceReport now returns zero for row totals when in cumulative or historical mode (#329)
# 1.14.1 2019-03-20
- require easytest <0.3 to fix build issue
# 1.14 2019-03-01
- added:
transaction, [v]post*, balassert* constructors, for tests etc.
- renamed:
porigin -> poriginal
- refactored:
transaction balancing & balance assertion checking (#438)
# 1.13.1 (2019/02/02)
- stop depending on here to avoid haskell-src-meta/stackage blockage.
# 1.13 (2019/02/01)
- in Journal's jtxns field, forecasted txns are appended rather than prepended
- API changes:
added:
+setFullPrecision
+setMinimalPrecision
+expectParseStateOn
+embedFileRelative
+hereFileRelative
changed:
- amultiplier -> aismultiplier
- Amount fields reordered for clearer debug output
- tpreceding_comment_lines -> tprecedingcomment, reordered
- Hledger.Data.TransactionModifier.transactionModifierToFunction -> modifyTransactions
- Hledger.Read.Common.applyTransactionModifiers -> Hledger.Data.Journal.journalModifyTransactions
- HelpTemplate -> CommandDoc
# 1.12 (2018/12/02)
- switch to megaparsec 7 (Alex Chen)
We now track the stack of include files in Journal ourselves, since
megaparsec dropped this feature.
- add 'ExceptT' layer to our parser monad again (Alex Chen)
We previously had a parser type, 'type ErroringJournalParser = ExceptT
String ...' for throwing parse errors without allowing further
backtracking. This parser type was removed under the assumption that it
would be possible to write our parser without this capability. However,
after a hairy backtracking bug, we would now prefer to have the option to
prevent backtracking.
- Define a 'FinalParseError' type specifically for the 'ExceptT' layer
- Any parse error can be raised as a "final" parse error
- Tracks the stack of include files for parser errors, anticipating the
removal of the tracking of stacks of include files in megaparsec 7
- Although a stack of include files is also tracked in the 'StateT
Journal' layer of the parser, it seems easier to guarantee correct
error messages in the 'ExceptT FinalParserError' layer
- This does not make the 'StateT Journal' stack redundant because the
'ExceptT FinalParseError' stack cannot be used to detect cycles of
include files
- more support for location-aware parse errors when re-parsing (Alex Chen)
- make 'includedirectivep' an 'ErroringJournalParser' (Alex Chen)
- drop Ord instance breaking GHC 8.6 build (Peter Simons)
- flip the arguments of (divide\|multiply)\[Mixed\]Amount
- showTransaction: fix a case showing multiple missing amounts
showTransaction could sometimes hide the last posting's amount even if
one of the other posting amounts was already implicit, producing invalid
transaction output.
- plog, plogAt: add missing newline
- split up journalFinalise, reorder journal finalisation steps (#893) (Jesse Rosenthal)
The `journalFinalise` function has been split up, allowing more granular
control.
- journalSetTime --> journalSetLastReadTime
- journalSetFilePath has been removed, use journalAddFile instead
# 1.11.1 (2018/10/06)
- add, lib: fix wrong transaction rendering in balance assertion errors
and when using the add command
# 1.11 (2018/9/30)
- compilation now works when locale is unset (#849)
- all unit tests have been converted from HUnit+test-framework to easytest
- doctests now run quicker by default, by skipping reloading between tests.
This can be disabled by passing --slow to the doctests test suite
executable.
- doctests test suite executable now supports --verbose, which shows
progress output as tests are run if doctest 0.16.0+ is installed
(and hopefully is harmless otherwise).
- doctests now support file pattern arguments, provide more informative output.
Limiting to just the file(s) you're interested can make doctest start
much quicker. With one big caveat: you can limit the starting files,
but it always imports and tests all other local files those import.
- a bunch of custom Show instances have been replaced with defaults,
for easier troubleshooting. These were sometimes obscuring
important details, eg in test failure output. Our new policy is:
stick with default derived Show instances as far as possible, but
when necessary adjust them to valid haskell syntax so pretty-show
can pretty-print them (eg when they contain Day values, cf
https://github.com/haskell/time/issues/101). By convention, when
fields are shown in less than full detail, and/or in double-quoted
pseudo syntax, we show a double period (..) in the output.
- Amount has a new Show instance. Amount's show instance hid
important details by default, and showing more details required
increasing the debug level, which was inconvenient. Now it has a
single show instance which shows more information, is fairly
compact, and is pretty-printable.
ghci> usd 1
OLD:
Amount {acommodity="$", aquantity=1.00, ..}
NEW:
Amount {acommodity = "$", aquantity = 1.00, aprice = NoPrice, astyle = AmountStyle "L False 2 Just '.' Nothing..", amultiplier = False}
MixedAmount's show instance is unchanged, but showMixedAmountDebug
is affected by this change:
ghci> putStrLn $ showMixedAmountDebug $ Mixed [usd 1]
OLD:
Mixed [Amount {acommodity="$", aquantity=1.00, aprice=, astyle=AmountStyle {ascommodityside = L, ascommodityspaced = False, asprecision = 2, asdecimalpoint = Just '.', asdigitgroups = Nothing}}]
NEW:
Mixed [Amount {acommodity="$", aquantity=1.00, aprice=, astyle=AmountStyle "L False 2 Just '.' Nothing.."}]
- Same-line & next-line comments of transactions, postings, etc.
are now parsed a bit more precisely (followingcommentp).
Previously, parsing no comment gave the same result as an empty
comment (a single newline); now it gives an empty string.\
Also, and perhaps as a consequence of the above, when there's no
same-line comment but there is a next-line comment, we'll insert an
empty first line, since otherwise next-line comments would get moved
up to the same line when rendered.
- Hledger.Utils.Test exports HasCallStack
- queryDateSpan, queryDateSpan' now intersect date AND'ed date spans
instead of unioning them, and docs are clearer.
- pushAccount -> pushDeclaredAccount
- jaccounts -> jdeclaredaccounts
- AutoTransaction.hs -> PeriodicTransaction.hs & TransactionModifier.hs
- Hledger.Utils.Debug helpers have been renamed/cleaned up
# 1.10 (2018/6/30)
- build cleanly with all supported GHC versions again (7.10 to 8.4)
- support/use latest base-compat (#794)
- support/require megaparsec 6.4+
- extensive refactoring and cleanup of parsers and related types and utilities
- readJournalFile(s) cleanup, these now use InputOpts
- doctests now run a bit faster (#802)
# 1.9.1 (2018/4/30)
- new generic PeriodicReport, and some report-related type aliases
- new BudgetReport
- make (readJournal\|tryReader)s?WithOpts the default api, dropping "WithOpts"
- automated postings and command line account aliases happen earlier
in journal processing (see hledger changelog)
# 1.9 (2018/3/31)
- support ghc 8.4, latest deps
- when the system text encoding is UTF-8, ignore any UTF-8 BOM prefix
found when reading files.
- CompoundBalanceReport amounts are now normally positive.
The bs/bse/cf/is commands now show normal income, liability and equity
balances as positive. Negative numbers now indicate a contra-balance
(eg an overdrawn checking account), a net loss, a negative net worth,
etc. This makes these reports more like conventional financial
statements, and easier to read and share with others. (experimental)
- splitSpan now returns no spans for an empty datespan
- don't count periodic/modifier txns in Journal debug output
- lib/ui/web/api: move embedded manual files to extra-source-files
- Use skipMany/skipSome for parsing spacenonewline (Moritz Kiefer)
This avoids allocating the list of space characters only to then
discard it.
- rename, clarify purpose of balanceReportFromMultiBalanceReport
- fix some hlint warnings
- add some easytest tests
# 1.5 (2017/12/31)
- -V/--value uses today's market prices by default, not those of last transaction date. #683, #648)
- csv: allow balance assignment (balance assertion only, no amount) in csv records (Nadrieril)
- journal: allow space as digit group separator character, #330 (Mykola Orliuk)
- journal: balance assertion errors now show line of failed assertion posting, #481 (Sam Jeeves)
- journal: better errors for directives, #402 (Mykola Orliuk)
- journal: better errors for included files, #660 (Mykola Orliuk)
- journal: commodity directives in parent files are inherited by included files, #487 (Mykola Orliuk)
- journal: commodity directives limits precision even after -B, #509 (Mykola Orliuk)
- journal: decimal point/digit group separator chars are now inferred from an applicable commodity directive or default commodity directive. #399, #487 (Mykola Orliuk)
- journal: numbers are parsed more strictly (Mykola Orliuk)
- journal: support Ledger-style automated postings, enabled with --auto flag (Dmitry Astapov)
- journal: support Ledger-style periodic transactions, enabled with --forecast flag (Dmitry Astapov)
- period expressions: fix "nth day of {week,month}", which could generate wrong intervals (Dmitry Astapov)
- period expressions: month names are now case-insensitive (Dmitry Astapov)
- period expressions: stricter checking for invalid expressions (Mykola Orliuk)
- period expressions: support "every 11th Nov" (Dmitry Astapov)
- period expressions: support "every 2nd Thursday of month" (Dmitry Astapov)
- period expressions: support "every Tuesday", short for "every <n>th day of week" (Dmitry Astapov)
- remove upper bounds on all but hledger* and base (experimental)
It's rare that my deps break their api or that newer versions must
be avoided, and very common that they release new versions which I
must tediously and promptly test and release hackage revisions for
or risk falling out of stackage. Trying it this way for a bit.
# 1.4 (2017/9/30)
- add readJournalFile\[s\]WithOpts, with simpler arguments and support
for detecting new transactions since the last read.
- query: add payee: and note: query terms, improve description/payee/note docs (Jakub Zárybnický, Simon Michael, #598, #608)
- journal, cli: make trailing whitespace significant in regex account aliases
Trailing whitespace in the replacement part of a regular expression
account alias is now significant. Eg, converting a parent account to
just an account name prefix: --alias '/:acct:/=:acct'
- timedot: allow a quantity of seconds, minutes, days, weeks, months
or years to be logged as Ns, Nm, Nd, Nw, Nmo, Ny
- csv: switch the order of generated postings, so account1 is first.
This simplifies things and facilitates future improvements.
- csv: show the "creating/using rules file" message only with --debug
- csv: fix multiple includes in one rules file
- csv: add "newest-first" rule for more robust same-day ordering
- deps: allow ansi-terminal 0.7
- deps: add missing parsec lower bound, possibly related to #596, fpco/stackage#2835
- deps: drop oldtime flag, require time 1.5+
- deps: remove ghc < 7.6 support, remove obsolete CPP conditionals
- deps: fix test suite with ghc 8.2
# 1.3.1 (2017/8/25)
- Fix a bug with -H showing nothing for empty periods (#583, Nicholas Niro)
This patch fixes a bug that happened when using the -H option on
a period without any transaction. Previously, the behavior was no
output at all even though it should have shown the previous ending balances
of past transactions. (This is similar to previously using -H with -E,
but with the extra advantage of not showing empty accounts)
- allow megaparsec 6 (#594)
- allow megaparsec-6.1 (Hans-Peter Deifel)
- fix test suite with Cabal 2 (#596)
# 1.3 (2017/6/30)
journal: The "uncleared" transaction/posting status, and associated UI flags
and keys, have been renamed to "unmarked" to remove ambiguity and
confusion. This means that we have dropped the `--uncleared` flag,
and our `-U` flag now matches only unmarked things and not pending
ones. See the issue and linked mail list discussion for more
background. (#564)
csv: assigning to the "balance" field name creates balance
assertions (#537, Dmitry Astapov).
csv: Doubled minus signs are handled more robustly (fixes #524, Nicolas Wavrant, Simon Michael)
Multiple "status:" query terms are now OR'd together. (#564)
deps: allow megaparsec 5.3.
# 1.2 (2017/3/31)
## journal format
A pipe character can optionally be used to delimit payee names in
transaction descriptions, for more accurate querying and pivoting by
payee. Eg, for a description like `payee name | additional notes`,
the two parts will be accessible as pseudo-fields/tags named `payee`
and `note`.
Some journal parse errors now show the range of lines involved, not just the first.
## ledger format
The experimental `ledger:` reader based on the WIP ledger4 project has
been disabled, reducing build dependencies.
## Misc
Fix a bug when tying the knot between postings and their parent transaction, reducing memory usage by about 10% (#483) (Mykola Orliuk)
Fix a few spaceleaks (#413) (Moritz Kiefer)
Add Ledger.Parse.Text to package.yaml, fixing a potential build failure.
Allow megaparsec 5.2 (#503)
Rename optserror -> usageError, consolidate with other error functions
# 1.1 (2016/12/31)
## journal format
- balance assignments are now supported (#438, #129, #157, #288)
This feature also brings a slight performance drop (\~5%);
optimisations welcome.
- also recognise `*.hledger` files as hledger journal format
## ledger format
- use ledger-parse from the ledger4 project as an alternate reader for C++ Ledger journals
The idea is that some day we might get better compatibility with Ledger files this way.
Right now this reader is not very useful and will be used only if you explicitly select it with a `ledger:` prefix.
It parses transaction dates, descriptions, accounts and amounts, and ignores everything else.
Amount parsing is delegated to hledger's journal parser, and malformed amounts might be silently ignored.
This adds at least some of the following as new dependencies for hledger-lib:
parsers, parsec, attoparsec, trifecta.
## misc
- update base lower bound to enforce GHC 7.10+
hledger-lib had a valid install plan with GHC 7.8, but currently requires GHC 7.10 to compile.
Now we require base 4.8+ everywhere to ensure the right GHC version at the start.
- Hledger.Read api cleanups
- rename dbgIO to dbg0IO, consistent with dbg0, and document a bug in dbg*IO
- make readJournalFiles \[f\] equivalent to readJournalFile f (#437)
- more general parser types enabling reuse outside of IO (#439)
# 1.0.1 (2016/10/27)
- allow megaparsec 5.0 or 5.1
# 1.0 (2016/10/26)
## timedot format
- new "timedot" format for retroactive/approximate time logging.
Timedot is a plain text format for logging dated, categorised
quantities (eg time), supported by hledger. It is convenient
for approximate and retroactive time logging, eg when the
real-time clock-in/out required with a timeclock file is too
precise or too interruptive. It can be formatted like a bar
chart, making clear at a glance where time was spent.
## timeclock format
- renamed "timelog" format to "timeclock", matching the emacs package
- sessions can no longer span file boundaries (unclocked-out
sessions will be auto-closed at the end of the file).
- transaction ids now count up rather than down (#394)
- timeclock files no longer support default year directives
- removed old code for appending timeclock transactions to journal transactions.
A holdover from the days when both were allowed in one file.
## csv format
- fix empty field assignment parsing, rule parse errors after megaparsec port (#407) (Hans-Peter Deifel)
## journal format
- journal files can now include timeclock or timedot files (#320)
(but not yet CSV files).
- fixed an issue with ordering of same-date transactions included from other files
- the "commodity" directive and "format" subdirective are now supported, allowing
full control of commodity style (#295) The commodity directive's
format subdirective can now be used to override the inferred
style for a commodity, eg to increase or decrease the
precision. This is at least a good workaround for #295.
- Ledger-style "apply account"/"end apply account" directives are now used to set a default parent account.
- the Ledger-style "account" directive is now accepted (and ignored).
- bracketed posting dates are more robust (#304)
Bracketed posting dates were fragile; they worked only if you
wrote full 10-character dates. Also some semantics were a bit
unclear. Now they should be robust, and have been documented
more clearly. This is a legacy undocumented Ledger syntax, but
it improves compatibility and might be preferable to the more
verbose "date:" tags if you write posting dates often (as I do).
Internally, bracketed posting dates are no longer considered to
be tags. Journal comment, tag, and posting date parsers have
been reworked, all with doctests.
- balance assertion failure messages are clearer
- with --debug=2, more detail about balance assertions is shown.
## misc
- file parsers have been ported from Parsec to Megaparsec \o/ (#289, #366) (Alexey Shmalko, Moritz Kiefer)
- most hledger types have been converted from String to Text, reducing memory usage by 30%+ on large files
- file parsers have been simplified for easier troubleshooting (#275).
The journal/timeclock/timedot parsers, instead of constructing
opaque journal update functions which are later applied to build
the journal, now construct the journal directly by modifying the
parser state. This is easier to understand and debug. It also
rules out the possibility of journal updates being a space
leak. (They weren't, in fact this change increased memory usage
slightly, but that has been addressed in other ways). The
ParsedJournal type alias has been added to distinguish
"being-parsed" journals and "finalised" journals.
- file format detection is more robust.
The Journal, Timelog and Timedot readers' detectors now check
each line in the sample data, not just the first one. I think the
sample data is only about 30 chars right now, but even so this
fixed a format detection issue I was seeing.
Also, we now always try parsing stdin as journal format (not just sometimes).
- all file formats now produce transaction ids, not just journal (#394)
- git clone of the hledger repo on windows now works (#345)
- added missing benchmark file (#342)
- our stack.yaml files are more compatible across stack versions (#300)
- use newer file-embed to fix ghci working directory dependence (<https://github.com/snoyberg/file-embed/issues/18>)
- report more accurate dates in account transaction report when postings have their own dates
(affects hledger-ui and hledger-web registers).
The newly-named "transaction register date" is the date to be
displayed for that transaction in a transaction register, for
some current account and filter query. It is either the
transaction date from the journal ("transaction general date"),
or if postings to the current account and matched by the
register's filter query have their own dates, the earliest of
those posting dates.
- simplify account transactions report's running total.
The account transactions report used for hledger-ui and -web
registers now gives either the "period total" or "historical
total", depending strictly on the --historical flag. It doesn't
try to indicate whether the historical total is the accurate
historical balance (which depends on the user's report query).
- reloading a file now preserves the effect of options, query arguments etc.
- reloading a journal should now reload all included files as well.
- the Hledger.Read.* modules have been reorganised for better reuse.
Hledger.Read.Utils has been renamed Hledger.Read.Common
and holds low-level parsers & utilities; high-level read
utilities are now in Hledger.Read.
- clarify amount display style canonicalisation code and terminology a bit.
Individual amounts still have styles; from these we derive
the standard "commodity styles". In user docs, we might call
these "commodity formats" since they can be controlled by the
"format" subdirective in journal files.
- Journal is now a monoid
- expandPath now throws a proper IO error
- more unit tests, start using doctest
0.27 (2015/10/30)
- The main hledger types now derive NFData, which makes it easier to
time things with criterion.
- Utils has been split up more.
- Utils.Regex: regular expression compilation has been memoized, and
memoizing versions of regexReplace\[CI\] have been added, since
compiling regular expressions every time seems to be quite
expensive (#244).
- Utils.String: strWidth is now aware of multi-line strings (#242).
- Read: parsers now use a consistent p suffix.
- New dependencies: deepseq, uglymemo.
- All the hledger packages' cabal files are now generated from
simpler, less redundant yaml files by hpack, in principle. In
practice, manual fixups are still needed until hpack gets better,
but it's still a win.
0.26 (2015/7/12)
- allow year parser to handle arbitrarily large years
- Journal's Show instance reported one too many accounts
- some cleanup of debug trace helpers
- tighten up some date and account name parsers (don't accept leading spaces; hadddocks)
- drop regexpr dependency
0.25.1 (2015/4/29)
- support/require base-compat >0.8 (#245)
0.25 (2015/4/7)
- GHC 7.10 compatibility (#239)
0.24.1 (2015/3/15)
- fix JournalReader "ctx" compilation warning
- add some type signatures in Utils to help make ghci-web
0.24 (2014/12/25)
- fix combineJournalUpdates folding order
- fix a regexReplaceCI bug
- fix a splitAtElement bug with adjacent separators
- mostly replace slow regexpr with regex-tdfa (fixes #189)
- use the modern Text.Parsec API
- allow transformers 0.4*
- regexReplace now supports backreferences
- Transactions now remember their parse location in the journal file
- export Regexp types, disambiguate CsvReader's similarly-named type
- export failIfInvalidMonth/Day (fixes #216)
- track the commodity of zero amounts when possible
(useful eg for hledger-web's multi-commodity charts)
- show posting dates in debug output
- more debug helpers
0.23.3 (2014/9/12)
- allow transformers 0.4*
0.23.2 (2014/5/8)
- postingsReport: also fix date sorting of displayed postings (#184)
0.23.1 (2014/5/7)
- postingsReport: with disordered journal entries, postings before the
report start date could get wrongly included. (#184)
0.23 (2014/5/1)
- orDatesFrom -> spanDefaultsFrom
0.22.2 (2014/4/16)
- display years before 1000 with four digits, not three
- avoid pretty-show to build with GHC < 7.4
- allow text 1.1, drop data-pprint to build with GHC 7.8.x
0.22.1 (2014/1/6) and older: see http://hledger.org/release-notes or doc/CHANGES.md.
Reference in New Issue View Git Blame Copy Permalink