# [PATCH v2 0/2] git-p4: improve formatting of numeric values 7 messages from 2021-12-17 to 2021-12-20. Participants: Joel Holdsworth, Joachim Kuebart, Luke Diamand, Junio C Hamano. Thread: https://gitlist.dev/t/57111 ## Joel Holdsworth, 2021-12-17 20:38 Subject: [PATCH v2 0/2] git-p4: improve formatting of numeric values Message-ID: <20211217203856.2339161-1-jholdsworth@nvidia.com> URL: https://gitlist.dev/e/20211217203856.2339161-1-jholdsworth%40nvidia.com ``` This patch-set contains patches designed to improvement the way in which the git-p4 script reports files sizes and percentage values in an effort to improve the readability of the logged messages. Joel Holdsworth (2): git-p4: print size values in appropriate units git-p4: show progress as an integer git-p4.py | 25 +++++++++++++++++++------ 1 file changed, 19 insertions(+), 6 deletions(-) -- 2.34.1 ``` ## Joel Holdsworth, 2021-12-17 20:38 Subject: [PATCH v2 1/2] git-p4: print size values in appropriate units Message-ID: <20211217203856.2339161-2-jholdsworth@nvidia.com> URL: https://gitlist.dev/e/20211217203856.2339161-2-jholdsworth%40nvidia.com In-Reply-To: <20211217203856.2339161-1-jholdsworth@nvidia.com> ``` The git-p4 script reports file sizes in various log messages. Previously, in each case the script would print them as the number of bytes divided by 1048576 i.e. the size in mebibytes, rounded down to an integer. This resulted in small files being described as having a size of "0 MB". This patch replaces the existing behaviour with a new helper function: format_size_human_readable, which takes a number of bytes (or any other quantity), and computes the appropriate prefix to use: none, Ki, Mi, Gi, Ti, Pi, Ei, Zi, Yi. For example, a size of 123456 will now be printed as "120.6 KiB" greatly improving the readability of the log output. Large valued prefixes such as pebi, exbi, zebi and yobi are included for completeness, though they not expected to appear in any real-world Perforce repository! Signed-off-by: Joel Holdsworth --- This second version of the patch makes the following changes: * Corrects an issue with the messages being suffixed with an spurious "MB" text. * Makes the units of format_size_human_readable hard-coded to bytes. git-p4.py | 22 +++++++++++++++++----- 1 file changed, 17 insertions(+), 5 deletions(-) diff --git a/git-p4.py b/git-p4.py index 2b4500226a..4d8a249b85 100755 --- a/git-p4.py +++ b/git-p4.py @@ -56,6 +56,16 @@ p4_access_checked = False +def format_size_human_readable(num): + """ Returns a number of units (typically bytes) formatted as a human-readable + string. + """ + for unit in ["", "Ki", "Mi", "Gi", "Ti", "Pi", "Ei", "Zi"]: + if abs(num) < 1024.0: + return "{:3.1f} {}B".format(num, unit) + num /= 1024.0 + return "{:.1f} Yi{}B".format(num) + def p4_build_cmd(cmd): """Build a suitable p4 command line. @@ -2966,7 +2976,8 @@ def streamOneP4File(self, file, contents): size = int(self.stream_file['fileSize']) else: size = 0 # deleted files don't get a fileSize apparently - sys.stdout.write('\r%s --> %s (%i MB)\n' % (file_path, relPath, size/1024/1024)) + sys.stdout.write('\r%s --> %s (%s)\n' % ( + file_path, relPath, format_size_human_readable(size))) sys.stdout.flush() (type_base, type_mods) = split_p4_type(file["type"]) @@ -3064,9 +3075,8 @@ def streamP4FilesCb(self, marshalled): if not err and 'fileSize' in self.stream_file: required_bytes = int((4 * int(self.stream_file["fileSize"])) - calcDiskFree()) if required_bytes > 0: - err = 'Not enough space left on %s! Free at least %i MB.' % ( - os.getcwd(), required_bytes/1024/1024 - ) + err = 'Not enough space left on %s! Free at least %s.' % ( + os.getcwd(), format_size_human_readable(required_bytes)) if err: f = None @@ -3110,7 +3120,9 @@ def streamP4FilesCb(self, marshalled): size = int(self.stream_file["fileSize"]) if size > 0: progress = 100*self.stream_file['streamContentSize']/size - sys.stdout.write('\r%s %d%% (%i MB)' % (self.stream_file['depotFile'], progress, int(size/1024/1024))) + sys.stdout.write('\r%s %d%% (%s)' % ( + self.stream_file['depotFile'], progress, + format_size_human_readable(required_bytes))) sys.stdout.flush() self.stream_have_file_info = True -- 2.34.1 ``` ## Joel Holdsworth, 2021-12-17 20:38 Subject: [PATCH v2 2/2] git-p4: show progress as an integer Message-ID: <20211217203856.2339161-3-jholdsworth@nvidia.com> URL: https://gitlist.dev/e/20211217203856.2339161-3-jholdsworth%40nvidia.com In-Reply-To: <20211217203856.2339161-1-jholdsworth@nvidia.com> ``` When importing files from Perforce, git-p4 periodically logs the progress of file transfers as a percentage. However, the value is printed as a float with an excessive number of decimal places. For example a typical update might contain the following message: Importing revision 12345 (26.199617677553135%) This patch simply rounds the value down to the nearest integer percentage value, greatly improving readability. Signed-off-by: Joel Holdsworth --- git-p4.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/git-p4.py b/git-p4.py index 4d8a249b85..3c621a6efd 100755 --- a/git-p4.py +++ b/git-p4.py @@ -3635,7 +3635,8 @@ def importChanges(self, changes, origin_revision=0): self.updateOptionDict(description) if not self.silent: - sys.stdout.write("\rImporting revision %s (%s%%)" % (change, cnt * 100 / len(changes))) + sys.stdout.write("\rImporting revision %s (%d%%)" % ( + change, (cnt * 100) // len(changes))) sys.stdout.flush() cnt = cnt + 1 -- 2.34.1 ``` ## Joachim Kuebart, 2021-12-18 07:44 Subject: Re: [PATCH v2 1/2] git-p4: print size values in appropriate units Message-ID: <39F3A9A9-AE61-4C2E-AA66-6B072634D91E@gmail.com> URL: https://gitlist.dev/e/39F3A9A9-AE61-4C2E-AA66-6B072634D91E%40gmail.com In-Reply-To: <20211217203856.2339161-2-jholdsworth@nvidia.com> ``` > On 17 Dec 2021, at 21:38, Joel Holdsworth wrote: > > git-p4.py | 22 +++++++++++++++++----- > 1 file changed, 17 insertions(+), 5 deletions(-) > > diff --git a/git-p4.py b/git-p4.py > index 2b4500226a..4d8a249b85 100755 > --- a/git-p4.py > +++ b/git-p4.py > @@ -56,6 +56,16 @@ > > p4_access_checked = False > > +def format_size_human_readable(num): > + """ Returns a number of units (typically bytes) formatted as a human-readable > + string. > + """ > + for unit in ["", "Ki", "Mi", "Gi", "Ti", "Pi", "Ei", "Zi"]: > + if abs(num) < 1024.0: > + return "{:3.1f} {}B".format(num, unit) > + num /= 1024.0 > + return "{:.1f} Yi{}B".format(num) This now has an extra pair of braces. Cheers, Joachim ``` ## Luke Diamand, 2021-12-18 10:09 Subject: Re: [PATCH v2 1/2] git-p4: print size values in appropriate units Message-ID: URL: https://gitlist.dev/e/CAE5ih7_knaVsxYBkWrJO4%2BjURM02E98iL9juroahW0k%3DZF7Oiw%40mail.gmail.com In-Reply-To: <39F3A9A9-AE61-4C2E-AA66-6B072634D91E@gmail.com> ``` On Sat, 18 Dec 2021 at 07:44, Joachim Kuebart wrote: > > > > > On 17 Dec 2021, at 21:38, Joel Holdsworth wrote: > > > > git-p4.py | 22 +++++++++++++++++----- > > 1 file changed, 17 insertions(+), 5 deletions(-) > > > > diff --git a/git-p4.py b/git-p4.py > > index 2b4500226a..4d8a249b85 100755 > > --- a/git-p4.py > > +++ b/git-p4.py > > @@ -56,6 +56,16 @@ > > > > p4_access_checked = False > > > > +def format_size_human_readable(num): > > + """ Returns a number of units (typically bytes) formatted as a human-readable > > + string. > > + """ > > + for unit in ["", "Ki", "Mi", "Gi", "Ti", "Pi", "Ei", "Zi"]: > > + if abs(num) < 1024.0: > > + return "{:3.1f} {}B".format(num, unit) > > + num /= 1024.0 > > + return "{:.1f} Yi{}B".format(num) > > This now has an extra pair of braces. > > Cheers, > Joachim It also seems to add some slightly spurious floating point precision. For example, 1<<20 comes out as "1.0 MiB" and 50 bytes comes out as "50.0 B". There's actually an older python2->python3 conversion bug going on here. The code uses: Size / 1024 / 1024 With python2 that came out as an integer, which for bytes is what you want. For python3 it comes out as a float - it should have been converted to Size // 1024 // 1024. I guess Joel was trying to preserve that original bug, but really we should just fix it! Overall I think this would definitely be an improvement, just needs to drop the spurious precision (and fix the return for values beyond ZiB as noted by Joachim). Luke ``` ## Luke Diamand, 2021-12-18 10:12 Subject: Re: [PATCH v2 2/2] git-p4: show progress as an integer Message-ID: URL: https://gitlist.dev/e/CAE5ih7_uq2xos46tH_dR%2BgietA_5bhr3JzYNJOLmrjhuijy1Ng%40mail.gmail.com In-Reply-To: <20211217203856.2339161-3-jholdsworth@nvidia.com> ``` On Fri, 17 Dec 2021 at 20:39, Joel Holdsworth wrote: > > When importing files from Perforce, git-p4 periodically logs the > progress of file transfers as a percentage. However, the value is > printed as a float with an excessive number of decimal places. > > For example a typical update might contain the following message: > > Importing revision 12345 (26.199617677553135%) > > This patch simply rounds the value down to the nearest integer > percentage value, greatly improving readability. > > Signed-off-by: Joel Holdsworth > --- > git-p4.py | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/git-p4.py b/git-p4.py > index 4d8a249b85..3c621a6efd 100755 > --- a/git-p4.py > +++ b/git-p4.py > @@ -3635,7 +3635,8 @@ def importChanges(self, changes, origin_revision=0): > self.updateOptionDict(description) > > if not self.silent: > - sys.stdout.write("\rImporting revision %s (%s%%)" % (change, cnt * 100 / len(changes))) > + sys.stdout.write("\rImporting revision %s (%d%%)" % ( > + change, (cnt * 100) // len(changes))) > sys.stdout.flush() > cnt = cnt + 1 Thanks for fixing this - this is one of those small python2->python3 bugs that's still left over. Looks good to me, ack. Luke ``` ## Junio C Hamano, 2021-12-20 20:02 Subject: Re: [PATCH v2 2/2] git-p4: show progress as an integer Message-ID: URL: https://gitlist.dev/e/xmqqzgovujs2.fsf%40gitster.g In-Reply-To: ``` Luke Diamand writes: >> diff --git a/git-p4.py b/git-p4.py >> index 4d8a249b85..3c621a6efd 100755 >> --- a/git-p4.py >> +++ b/git-p4.py >> @@ -3635,7 +3635,8 @@ def importChanges(self, changes, origin_revision=0): >> self.updateOptionDict(description) >> >> if not self.silent: >> - sys.stdout.write("\rImporting revision %s (%s%%)" % (change, cnt * 100 / len(changes))) >> + sys.stdout.write("\rImporting revision %s (%d%%)" % ( >> + change, (cnt * 100) // len(changes))) >> sys.stdout.flush() >> cnt = cnt + 1 > > Thanks for fixing this - this is one of those small python2->python3 > bugs that's still left over. > > Looks good to me, ack. Thanks. I'll wait for an updated [1/2] but this is an independently good change that does not depend on it, so I'll take [2/2] alone for now. ```