== Background ==
Reporting from Xenial, but this is currently reproducible in Artsy too.
$ locale
LANG=fi_FI.UTF-8
LANGUAGE=fi:en
LC_CTYPE=fi_FI.UTF-8
LC_NUMERIC=”fi_FI.UTF-8″
LC_TIME=”fi_FI.UTF-8″
LC_COLLATE=fi_FI.UTF-8
LC_MONETARY=”fi_FI.UTF-8″
LC_MESSAGES=fi_FI.UTF-8
LC_PAPER=”fi_FI.UTF-8″
LC_NAME=”fi_FI.UTF-8″
LC_ADDRESS=”fi_FI.UTF-8″
LC_TELEPHONE=”fi_FI.UTF-8″
LC_MEASUREMENT=”fi_FI.UTF-8″
LC_IDENTIFICATION=”fi_FI.UTF-8″
LC_ALL=
== Test case ==
$ git init test
$ cd test
$ echo ”foo” > bär
$ git add [hit tab key]
$ git add \”b\\303\\244r\” [hit enter]
== What happens ==
fatal: pathspec ’”b\303\244r”’ did not match any files
== What I expect to happen ==
For the filename to be correctly completed, like with ls:
$ mkdir test2
$ cd test2
$ echo ”foo” > bär
$ ls [hit tab key]
$ ls bär [hit enter]
bär
For optimizing files in-place with jpegtran, especially from scripts and when dealing with lots of pictures, it’s handy to be able to specify input file as -outfile.
But there’s a catch:
Steps to reproduce:
0. Have a large JPEG file, or, alternatively, somewhat slow CPU
1. `jpegtran -optimize -copy all -perfect -outfile large.jpg large.jpg`
2. Hit Ctrl-C before the command finishes
Result:
You now have a broken large.jpg with only part, if any, of the image data remaining.
What I expect to happen:
To have large.jpg as it was before I invoked jpegtran.
Workarounds:
Obviously the traditional workaround of specifying an intermediate temporary output file, then replacing the original with the temporary file only once jpegtran has finished.
Depending on use case, using `fmt -w 80` instead of fold may do for a workaround.
Technical details
- YOURLS version: master (at e36ec69)
- PHP version: 5.6.30
- I’m running Yourls on shared hosting and they have php5-geoip installed
Bug Summary
After pulling recent changes (up to e36ec69) today, I’ve hit what appears to be a regression of issue #359: I get a blank page and a ”Cannot redeclare geoip_country_code_by_name()” (in includes/geo/geoip.inc on line 1602) error in my log.
I can work around the issue by checking out d5a5522, which is the last commit prior to ”Update GeoIP” (96a159b).
I make no guarantees this doesn’t break something else, but for me it seems to work correctly if I make Feediron_Json::format() skip carriage returns:
$ git diff
diff --git a/Json.php b/Json.php
index c988962..fc7f554 100644
--- a/Json.php
+++ b/Json.php
@@ -8,6 +8,7 @@ class Feediron_Json{
$strLen = strlen($json);
$indentStr = ' ';
$newLine = "\n";
+ $cReturn = "\r";
$prevChar = '';
$outOfQuotes = true;
$currentline = 0;
@@ -30,6 +31,9 @@ class Feediron_Json{
$currentline++;
continue;
}
+ if($char == $cReturn){
+ continue;
+ }
if($char == ' ' && $outOfQuotes){
continue;
}
(Using PHP 5.6.30 here.)
@davesteele Yup, installs without a hitch now. Much appreciated!
Was there a regression or is gtk-update-icon-cache now a hard dependency? I’m getting this in Ubuntu 16.04 (when using the PPA).
The following packages have unmet dependencies:
gnome-gmail : Depends: gtk-update-icon-cache but it is not installable
$ zsync http://cdimage.ubuntu.com/ubuntu-server/daily/current/zesty-server-amd64.iso.zsync
#################### 100.0% 144.7 kBps DONE
No relevent local data found - I will be downloading the whole file. If that's not what you want, CTRL-C out. You should specify the local file is the old version of the file to download with -i (you might have to decompress it with gzip -d first). Or perhaps you just have no data that helps download the file
What I *think* it’s trying to say is something like: ”No relevant local data was found, and therefore the entire file will be downloaded. If that’s not what you want, use Ctrl-C to abort. If you do have local data, you should specify the location with -i. If the local data is compressed, decompress it first.”
Also, the ”verifying download…checksum matches OK” message (after downloading is finished) could do with capitalization and whitespace: ”Verifying download… Checksum matches OK.”
Redefining TimeoutIdle or TimeoutNoTransfer in a configuration file inside /etc/proftpd/conf.d/, after they’ve already been defined in /etc/proftpd/proftpd.conf, fails to change those values for the daemon — that is, after restarting the daemon, the values for those parameters remain as those set in /etc/proftpd/proftpd.conf.
At least TimeoutIdle or TimeoutNoTransfer are affected. TimeoutSession and TimeoutLogin, which (in stock 16.04) haven’t been defined in the master configuration file can be set effectively, but setting them first in the master file, then again to different values inside conf.d reveals them to be equally affected.
On the other hand, (at least) ServerName *can* be overridden in the custom configuration file, so the handling of different configuration knobs in this regard seems to be inconsistent.
Steps to reproduce:
1. Create a /etc/proftpd/conf.d/proftpd.conf with the following lines:
TimeoutIdle 10
TimeoutNoTransfer 10
2. Restart proftpd
3. Log into the server, wait 11 seconds
What I expect to happen:
To get kicked out of the server.
What happens instead:
I’m allowed to linger on the server, presumably for the 600 seconds defined for TimeoutNoTransfer in /etc/proftpd/proftpd.conf.
The Include directive, when given a directory as parameter (such as /etc/proftpd/conf.d/, as in the stock /etc/proftpd/proftpd.conf), causes all files in said directory to be read, not just ones ending in .conf. This causes problems if, for instance, I’m using vim to edit a file in the included directory while the proftpd service is being (re)started: vim stores a .swp file in the same directory, and proftpd may fail to start with ”fatal: unknown configuration directive” when it tries to parse the .swp file.