there is no real reason to maintain the order of files, so we can just
use the sorted keys of ws_lines.

Change-Id: Id973fcf1fec6995a931dee665303fb58074bc320
Reviewed-by: Orgad Shaneh <orgads@gmail.com>

Change-Id: I999434f1953896d0bd0d2328176df65b05d63855
Reviewed-by: Orgad Shaneh <orgads@gmail.com>

it actually relies on $_ anyway

Change-Id: I17f48f1d50c20936ffd27a06bfb826be7c3bdd7f
Reviewed-by: Orgad Shaneh <orgads@gmail.com>

they are generated, and we need to check them in.

Change-Id: Ife8bd4a820c5a29ac9e28ce6af102d30577df4cb
Reviewed-by: Orgad Shaneh <orgads@gmail.com>

xcode always omits it.

Change-Id: I37b1e799d08d595aee4c227f9ed7af3eca69a6b8
Reviewed-by: Orgad Shaneh <orgads@gmail.com>

Change-Id: I6ad9c682697e80e175cb19d67d3167c1756cd311
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>

Submodules that have no changes actually show up with space in front of
the sha-1 which would mess up the logic when splitting by spaces.

Change-Id: I15506c6e3780efb29d885ff456ecfac60526c03d
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>

Change-Id: Ia8fb489715e2ba0f0c29ad76ada20800a1ab5c54
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>

- Added LICENSE.LGPLv21, LICENSE.LGPLv3 & LICENSE.GPLv2

Change-Id: If5b04ddf25095ae1eb59e25361c88e570bb2b4df
Reviewed-by: Antti Kokko <antti.kokko@digia.com>
Reviewed-by: Tony Sarajärvi <tony.sarajarvi@digia.com>

These files are only edited in Xcode in practice, which uses tabs.

Change-Id: Id13d2dc6c40955288041aeaee37282d41e9026f3
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>

Change-Id: I3f12d4ce0653b31ca397167a40303b764bdf7472
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

i considered making --draft the default. however, i didn't like the
change of default behavior.

then it occurred to me that automatically detecting WIP subjects like
the sanity bot does would actually fit our current workflow pretty
neatly.
we are looking only at the last commit, as that's likely to be the one
which is WIP. also, in "unbased" push mode, determining the start of the
range to look at would cause quite some overhead.

Change-Id: I70f6f397693f600b6a5d5791f04d9cae12fa8d02
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

at the moment the primary effect of this is ensuring early that we are,
in fact, inside a git clone.

Change-Id: I265631a0f4a025b09556ccbb9bfd735973d66d78
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

Change-Id: Id49de348f8808a9786616e512636c08eb4d6ef6d
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

properly quote where necessary and space-separate, instead of
comma-separating.

Change-Id: Ibf2e638d92591700c3a26f3402f306f5995552ad
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

Change-Id: Ia6b72ca42e2493842e228269728be46dd2eaf1df
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

Change-Id: I4d210bd3c896bb5d1cb13b4d38a08bfb317f4150
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

object-orientation just adds another layer of obfuscation to perl code,
and really doesn't pay off for such short scripts.

Change-Id: Ibe620839cea9b2cf0b2ff56257e13071535bd3a8
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>

Change-Id: I729dc9f37f604d7a7ed97ddb7395080b3afa3b3e
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

Change-Id: I019707fa3ca050fc9da8fb95b00766f661447223
Reviewed-by: Orgad Shaneh <orgads@gmail.com>

Change-Id: Ia5ca3a45702c52895f084f7f4bbe57ddb673b5ac
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>

Change-Id: I67dc5f1a1febeee8465abf34d0cb35bde1a868b9
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>

it's supposed to match a literal dot

Change-Id: I4bee3571a4239ec08a14ed18e409f97b4455f928
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

Change-Id: I2f3feab4a8396d73c8e3313731427061534536a4
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

people sometimes get creative

Change-Id: Idcb52ca5a4cc0561d80f6b8c74010d701c854dc0
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

otherwise we complain about the missing newline at the end, which is
pointless (arguably, git is being stupid here in the first place, but
whatever).

Change-Id: I51bc49825c2a74b26691927660a6771a3925af7d
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

it is expected to be a [ChangeLog] tag, not a ChangeLog footer.

Change-Id: Iaf5bb09976ecfbd4d7cbde9eaf303f682aa14f58
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Change-Id: I223eaf32782b2c0d2de3a815e0968706bfff1bdb
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>

urls to reviews should be replaced by sha1s of merged commits.
urls to jira tasks should be replaced by task-number footers.

the rationale is keeping the history relocatable, and somewhat
independent from the infrastructure's software.
also, urls are pretty noisy.

Change-Id: I4b2c9ded8cbc02ca96daa4a2ff92c98c0a4b05f9
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

people use them to refer to related commits. they should use stable
sha1s of merged commits instead.

Change-Id: Ic58aa875aa978dc76948521266741258fd8c94aa
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

they are completely unnecessary and make things look messy.

the algorithm considers cherry-pick lines as footers, following
convention. it even supports multi-line cherry-picks, such as those i
sometimes assemble when squashing multiple commits.

this change will cause a lot of complaints, because almost everybody
appears to be using a broken commit-msg hook. the one coming from our
gerrit is ok, as is the copy in qtrepotools.

Change-Id: I845ba16a70960beb2ce042b2eaddf22b17edcc25
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

it would be conceivable to integrate this with the bot driving this
script, so the message would be appended only to the first patchset in a
change, or even only to the first complaint (within a while) for any
particular contributor.

Change-Id: I5a7e3265dea8a3e24649d223fdd6e96c886eebc1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Change-Id: I6aff5775a2a9f9c685ba4121557e4c54a469304b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

Change-Id: Ifdeb8782bcc47730514eca1fb2f2a481c78484c7
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

Change-Id: I83d33248a0f9f17e2d44f841064e5280110c006b
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>

this was forgotten. it should be analogous to the rev-by check.

Change-Id: I82f35c13a9e17b03b196533e73aec8f240444265
Reviewed-by: Sergio Ahumada <sahumada@blackberry.com>

they are redundant with the canonical Task-number footers and steal
precious space in the summary line.

Change-Id: I339c450d6322a7c403b097d880bff10c8167d973
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

some people think a summary like "revert commit 124fa614423" is a good
idea. it's not - without researching the history, it is meaningless.

Change-Id: I6de02d911073aa2982ee05a6804db370a05d49e5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

we don't need them, as contributors do a pre-emptive sign-off on all their
contributions by agreeing with the CLA.

Change-Id: Ibab565370e00800dde7fddaa10a7aa21c332037f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

unlike previously thought, gerrit won't eliminate known bogus footers,
so they would stay behind. consequently, complain about them loudly.

note that the footers are just fine above a "(cherry picked from ...)"
line, as we want to preserve the original approvals.

Change-Id: I7856810adb3e3b3e91e539eb626375dd32648460
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>