git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Philip McGraw <Philip.McGraw@bentley.com>
To: "ahippo@yandex.ru" <ahippo@yandex.ru>
Cc: "git@vger.kernel.org" <git@vger.kernel.org>,
	"luke@diamand.org" <luke@diamand.org>
Subject: RE: [PATCH] git-p4: close temporary file before removing
Date: Thu, 1 Aug 2019 15:30:18 +0000	[thread overview]
Message-ID: <BL0PR1901MB2097141B412696422CE957F4FFDE0@BL0PR1901MB2097.namprd19.prod.outlook.com> (raw)
In-Reply-To: <2717551564623283@vla1-822b1b47a947.qloud-c.yandex.net>

> From: Andrey <ahippo@yandex.ru>
> Sent: Wednesday, 31 July, 2019 21:35
> To: Philip McGraw <Philip.McGraw@bentley.com>
> Cc: git@vger.kernel.org; luke@diamand.org
> Subject: Re: [PATCH] git-p4: close temporary file before removing
> 
> 31.07.2019, 17:52, "Philip McGraw" <philip.mcgraw@bentley.com>:
> > 2019.07.31 10:09 Andrey <ahippo@yandex.ru>
> >> 31.07.2019, 09:53, "Philip McGraw" <philip.mcgraw@bentley.com>:
> >>>>   30.07.2019, 13:37, "Philip McGraw" <philip.mcgraw@bentley.com>:
> >>>>   > python os.remove() throws exceptions on Windows platform when
> attempting
> >>>>   > to remove file while it is still open. Need to grab filename
> while file open,
> >>>>   > close file handle, then remove by name. Apparently other
> platforms are more
> >>>>   > permissive of removing files while busy.
> >>>>   > reference:
> >>>>   > ---
> >>>>   >  git-p4.py | 4 +++-
> >>>>   >  1 file changed, 3 insertions(+), 1 deletion(-)
> >>>>   >
> >>>>   > diff --git a/git-p4.py b/git-p4.py
> >>>>   > index c71a6832e2..6b9d2a8317 100755
> >>>>   > --- a/git-p4.py
> >>>>   > +++ b/git-p4.py
> >>>>   > @@ -1161,12 +1161,14 @@ def exceedsLargeFileThreshold(self,
> relPath, contents):
> >>>>   >                  return False
> >>>>   >              contentTempFile = self.generateTempFile(contents)
> >>>>   >              compressedContentFile =
> tempfile.NamedTemporaryFile(prefix='git-p4-large-file', delete=False)
> >>>>   > + compressedContentFileName = compressedContentFile.name
> >>>>   >              zf = zipfile.ZipFile(compressedContentFile.name,
> mode='w')
> >>>>   >              zf.write(contentTempFile,
> compress_type=zipfile.ZIP_DEFLATED)
> >>>>   >              zf.close()
> >>>>   >              compressedContentsSize =
> zf.infolist()[0].compress_size
> >>>>   >              os.remove(contentTempFile)
> >>>>   > - os.remove(compressedContentFile.name)
> >>>>   > + compressedContentFile.close()
> >>>>   > + os.remove(compressedContentFileName)
> >>>>
> >>>>   I'm not sure why NamedTemporaryFile() is called with delete=False
> above,
> >>>>   but it appears to me that it can have delete=True instead,
> >>>>   so that there is no need to call os.remove() explicitly
> >>>>   and thus worry about remove vs close ordering at all.
> >>>>
> >>>>   >              if compressedContentsSize > gitConfigInt('git-
> p4.largeFileCompressedThreshold'):
> >>>>   >                  return True
> >>>>   >          return False
> >>>>   > --
> >>>>   > 2.21.0.windows.1
> >>>>
> >>>>   Thank you,
> >>>>   Andrey.
> >>>
> >>>  Thanks Andrey; simpler is certainly better! I will test and re-submit
> v2 of patch with that approach.
> >>
> >> Thank you, that would be great!
> >>
> >> --
> >> Andrey.
> >
> > Unfortunately it wasn't as simple it seemed: upon testing with only
> changing delete=True,
> > found that the problem was not solved. Upon further debugging,
> recoded/refactored slightly adding
> > allocateTempFileName() locally scoped function to try to clarify how the
> NamedTemporaryFile()
> > was actually being used.
> >
> > We can't depend on the delete-on-close because the NamedTemporaryFile()
> is merely allocating
> > a temporary name for real use by the zipfile open-for-write which fails
> (on Windows) if file
> > was not explicitly closed first.
> 
> Oh, sorry for misguiding you!
> I didn't think of this aspect.

No worries! I probably just misunderstood the implementation of your idea.
> 
> > Hopefully the new patch
> (https://urldefense.proofpoint.com/v2/url?u=https-
> 3A__github.com_gitgitgadget_git_pull_301&d=DwIDaQ&c=hmGTLOph1qd_VnCqj81HzE
> WkDaxmYdIWRBdoFggzhj8&r=b0ikFMJGw7xxhF3yjexiWJpLuNxlAh1SvUDuUJ-
> pHmE&m=1jGOrV_I1Mg5ajkJ7yFEcNlyLnD6zYNXqXB9Z5SIPyE&s=TdT4WHyQCk5WZty_CvajH
> XgrZJbmIOl1gbMcngmjmAs&e= ) will make this more clear.
> 
> The new changeset looks good to me.
> (I'll post a reply in that thread too)
> 
> > Open to other suggestions if still not clear.
> 
> Just as a thought, ZipFile() can take a file-like object instead of a file
> name,
> so can be passed the NamedTemporaryFile() object directly instead of its
> file name.
> This should hopefully avoid double-open issue on Windows.

Another excellent idea that minimizes changes.  I am testing this approach
now and will submit v3 of the patch soon.
> 
> However, I'm good with your allocateTempFileName() changeset,
> so it's up to you to give it a try or not.
> 
> > Thanks again,
> > Philip
> 
> Thank you,
> Andrey.

Thanks,
Philip

  reply	other threads:[~2019-08-01 15:33 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-07-30 17:37 [PATCH] git-p4: close temporary file before removing Philip McGraw
2019-07-31  1:48 ` Andrey
2019-07-31 13:53   ` Philip McGraw
2019-07-31 14:09     ` Andrey
2019-07-31 21:51       ` Philip McGraw
2019-08-01  1:34         ` Andrey
2019-08-01 15:30           ` Philip McGraw [this message]
2019-08-02  3:50             ` Andrey

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: http://vger.kernel.org/majordomo-info.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=BL0PR1901MB2097141B412696422CE957F4FFDE0@BL0PR1901MB2097.namprd19.prod.outlook.com \
    --to=philip.mcgraw@bentley.com \
    --cc=ahippo@yandex.ru \
    --cc=git@vger.kernel.org \
    --cc=luke@diamand.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://80x24.org/mirrors/git.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).