Skip to content
Commits on Jan 6, 2016
  1. Version bump 0.1.3-dev

    committed Jan 7, 2016
  2. Update CHANGES.txt

    committed Jan 6, 2016
Commits on Jan 4, 2016
  1. @zerkms

    Proportional scale for the speaker overlay

    zerkms committed Jan 4, 2016
    Addresses #25
    
    The width now is fixed to be 320px and the height is proportionally calculated. Then the overlay offset is calculated based on that proportionally scaled height so that the overlay is snapped to the right bottom corner as before.
Commits on Sep 28, 2015
  1. Add missing entries in CHANGES.txt

    committed Sep 28, 2015
  2. Version bump 0.1.2-dev

    committed Sep 28, 2015
Commits on Sep 19, 2015
Commits on May 13, 2015
  1. Version bump 0.1.1-dev

    committed May 13, 2015
  2. Fix some encoding issues

    committed May 13, 2015
Commits on Mar 16, 2014
  1. Better error messages

    committed Mar 16, 2014
  2. Fix encoding issue.

    committed Mar 16, 2014
Commits on Jan 6, 2014
  1. Code and architecture cleanup.

    committed Jan 6, 2014
    Try to have a better module split and avoid useless classes.
  2. Fix documentation: DownloadFailedException has been renamed in Downlo…

    committed Jan 6, 2014
    …adError long time ago
  3. Better --overwrite handling.

    committed Jan 6, 2014
    It makes no sense wasting time and bandwidth if the file already exist.
    It is better to check if the file exist before doing anything
Commits on Jan 5, 2014
  1. Add docs/ submodule.

    committed Jan 5, 2014
Commits on Jan 4, 2014
  1. Add Python 3 support.

    committed Jan 4, 2014
    infoqscraper is now compatible with Python 2 and Python 3 (>= 2.6).
    Comptability is ensured by the six module.
    
    Currently only Python 2.7 and Python 3.3 have be tested. I plan to
    use travis to run a wider test campaign.
  2. Do some code cleanup in the downloader.

    committed Jan 4, 2014
    Removed some useless legacy code and fixed documentation.
  3. Get rid of the PIL dependency

    committed Jan 4, 2014
    PIL was used to convert PNG into JPEG files to feed ffmpeg because older releases
    of ffmpeg had issue with PNG.
    
    It seems now OK to accept PNG too. Affected user can download a static linked
    version of ffmpeg as a workaround.
    
    Removing PIL is important because it prevented the port of infoqscraper to python 3.
  4. Add several output type

    committed Jan 4, 2014
    The legacy format is still supported since it is compitable with quite old versions
    of ffmpeg.
    
    Add two new formats: h264 and h264_overlay. They both require a recent version of ffmpeg.
Commits on Jan 1, 2014
  1. version bump

    committed Jan 1, 2014
  2. Add few more files to .gitignore

    committed Jan 1, 2014
Commits on Dec 26, 2013
  1. Fix #21: Compatibility issue with ffmpeg 0.8

    committed Dec 26, 2013
    In 0.0.7 --overwrite has been introduced. It adds "-y" or "-n" to the
    ffmpeg command to allow to overwrite an already existing file or exit
    with error otherwise.
    
    Unfortunately this option has been added quite recently to ffmpeg and
    introduce a regression on Ubuntu 12.04, 13.10 and most likely lot of
    other stable distros. This patch use "-y" and replace "-n" by a manual
    check. A race condition can still occur but it is good enough.
  2. Remove --cache option from "presentation download"

    committed Dec 26, 2013
    It is redundant with "infoqscraper -c" and was not used.
Commits on Dec 9, 2013
  1. Version bump

    committed Dec 9, 2013
  2. Fix #16: Handle already existing file

    committed Dec 9, 2013
    Cleanly exit if the output file already exists.
    Add --overwrite/-y option to overwrite the existing file.
    
    Thanks to Lee Hanxue <hanxue.developer@gmail.com> for
    having reported the issue and provided a patch.
Commits on Dec 5, 2013
Commits on Dec 4, 2013
  1. @KevinGreene

    Fixed Regex for URLS

    KevinGreene committed Dec 4, 2013
    The old regex wouldn't match several characters at once, just the first one, so I added a + inside the parentheses.
    
    I also got rid of the $, as it seemed unnecessary. Things like a trailing slash, or other miscellaneous params added to the URL could ruin that. In order to make sure we only got the id, I also added # and ?, which are two other common url patterns that we wouldn't want to capture anything after.
    
    Other alternatives: 
    
    If you just want to get rid of a possible trailing slash: 
    
    ```
    mo = re.search("^https?://www.infoq.com/presentations/([^/]+)/?$", name) 
    ```
    
    If you want to assume that the id will always be alphanumeric + hyphens,
    
    ```
    mo = re.search("^https?://www.infoq.com/presentations/([0-9A-za-z-]+)", name) 
    ```
  2. @bytehead

    Added installation notes for OS X

    bytehead committed with Nov 30, 2013
  3. @bytehead
Commits on Oct 18, 2013
  1. version bump

    committed Oct 18, 2013
Something went wrong with that request. Please try again.