On 25.03.2015 22:04, Branko Čibej wrote: > On 25.03.2015 21:30, Branko Čibej wrote: >> On 25.03.2015 20:35, Dmitriy Setrakyan wrote: >>> On Wed, Mar 25, 2015 at 12:29 PM, Branko Čibej <[email protected]> wrote: >>> >>>> On 25.03.2015 20:11, Dmitriy Setrakyan wrote: >>>>> On Wed, Mar 25, 2015 at 11:55 AM, Branko Čibej <[email protected]> wrote: >>>>> >>>>>> On 25.03.2015 19:07, Dmitriy Setrakyan wrote: >>>>>>> Brane, >>>>>>> >>>>>>> The wrong download checksum issue has been addressed (I think). Please >>>>>>> double check me. The reason it happened was that I had two identical >>>>>>> folders with different names sitting next to each other and grabbed the >>>>>>> wrong one. The only difference was the name of the archive. (I should >>>>>>> really stop working after midnight :) >>>>>> Should be publishing exactly what was voted on, not even re-zipping, >>>>>> IMO. This is one of the reasons why it's a good idea to have the whole >>>>>> release process scripted; no manual archiving, copying, etc. anywhere. >>>>>> >>>>> This should be the case now. >>>> Out of interest, where do the Google Analytics reports go to? How can a >>>> lowly mentor have a look at the stats? >>>> >>> I have been reusing the GridGain one (mainly out of laziness). We should >>> setup the one for Ignite. I will work on it. >> Heh. You used the same element 'id' attribute twice on the download >> page; that's not valid HTML :) >> >> I'm fixing this ... > OK, the downloads page now points people at mirrors. Note that no all > mirrors are currently up-to-date; there's been less than 24 hours since > the last change on the site. That'll fix itself in due time. > > The magic that's happening here is as follows: > > * The 'Download' link points to the file 'downloads.html' > * This file causes the SSI module to execute download.cgi > * download.cgi finds the closest mirror, parses download.html, > interpolates the mirror info there and returns the new contents > in-place.
Actually, that's 'download.html -> download-mirrors.cgi -> download-mirrors.html' now; much easier to parse if it's not a one-letter difference. -- Brane
