bzr branch
http://darksoft.org/webbzr/adei/trunk
286
by Suren A. Chilingaryan
Update of setups |
1 |
Important
|
2 |
=========
|
|
3 |
! downloads_check.php - doesn't handle correctly failed downloads. We should |
|
4 |
do something about it. |
|
5 |
! For MSSQL column and table quotes should be changed from " to []. Yes, for |
|
6 |
this we should introduce opening and closing quote :( |
|
7 |
! Prevent entering custom plot mode if not supported |
|
8 |
||
9 |
Cache Engine
|
|
10 |
============
|
|
11 |
- We are currently just skipping databases which are failed to connect to the
|
|
12 |
data source (request.pph/GetSources/catch block). In case if we are caching the |
|
13 |
data this prevents administrator from getting information about the problems. |
|
14 |
- If we change use_subsecond flag without dropping the cache, the ADEI will
|
|
15 |
report channel inconsistency (not clarifying the issue). Since, we will get |
|
16 |
two extra columns in the caching table. The proper error message should be |
|
17 |
reported in this case. |
|
18 |
- At the moment, We can't really use non [\w\d_] symbols in gids if md5
|
|
19 |
postfixes are switched off. Otherwise, this symbols are stripped from the |
|
20 |
postfixes (for the sake of table names) and then the check for cached tables |
|
21 |
is failing (it reconstructs gid basing on the table names). Some, other things |
|
22 |
may fail as well. |
|
23 |
For the check, we can try to ignore the inappropriate symbols (only if md5 is |
|
24 |
switched off). So, the stripping should be performed by caching code. |
|
339
by Suren A. Chilingaryan
Update of ToDo lists |
25 |
- In the non-0 caching modes, the last interval is not displayed. Why cache is
|
26 |
not returning the last point? |
|
286
by Suren A. Chilingaryan
Update of setups |
27 |
|
312
by Suren A. Chilingaryan
Resizing of sidebar popups |
28 |
VIEWS
|
29 |
=====
|
|
30 |
- Support non-fitting timestamps in the scatter plot with resolution 0.
|
|
31 |
||
286
by Suren A. Chilingaryan
Update of setups |
32 |
Filters
|
33 |
=======
|
|
34 |
- While requesting last record (or potentially several last entries), it is
|
|
35 |
possible to get empty response if the last record while be rejected by the |
|
307
by Suren A. Chilingaryan
No data bug in scatterview |
36 |
configured filters. Iteratively increase limit? When fixed and may be |
37 |
universally used, shall we introduce a 'limit' option instead of making |
|
38 |
complex 'window' parsing? |
|
415
by Suren A. Chilingaryan
Synchronize ToDo lists with zweb |
39 |
- If we may want to filter the rows based on value of certain channel, we
|
40 |
may get problems with db_mask. It will be applied before the filter and |
|
41 |
only selected channels will pushed through. Using READER_Filter interface |
|
42 |
we may find the required columns, but will should ensure that the data |
|
43 |
required for filter is provided (introduce db_mask_required parameter of |
|
44 |
filter?) |
|
45 |
||
286
by Suren A. Chilingaryan
Update of setups |
46 |
|
47 |
Virtual
|
|
48 |
=======
|
|
49 |
- Support "__" in group names in SourceTree
|
|
50 |
||
51 |
Download Manager
|
|
52 |
===============
|
|
53 |
! Download manager does not respect "setup" |
|
54 |
! Show details of Downloads Manager button while not work if data is not |
|
55 |
from default setup. |
|
56 |
! download_manager.js is updating too often (due to adei.updater.Update() |
|
57 |
calls). This should be somehow fixed. As well it refers dlmanager.Request() |
|
58 |
which uses variable name set in correspndent module. This is completely |
|
59 |
insane and should be fixed. |
|
60 |
||
61 |
ZEUS Reader
|
|
62 |
===========
|
|
63 |
! Interlock between Full Backup and other calls in ZEUSReader. This is |
|
64 |
critical. If not prevented, the cache may start while configuration is |
|
65 |
partially dropped and not yet reloaded. This may result in rejecting |
|
66 |
some data vectors because of invalid number of records, etc. |
|
67 |
||
68 |
RRD Reader
|
|
69 |
==========
|
|
70 |
! Axes are not always cleaned when changing to different group, |
|
71 |
with RRD source |
|
72 |
- Do not list empty groups in RRDReader
|
|
73 |
||
74 |
Web Frontend
|
|
75 |
============
|
|
339
by Suren A. Chilingaryan
Update of ToDo lists |
76 |
- Rising memory consumption with Firefox (and I guess other browsers)
|
286
by Suren A. Chilingaryan
Update of setups |
77 |
- Passing away pointer to "this" in JS-constructors may be problematic (no
|
78 |
errors, but this will be an incorrect "this"). |
|
79 |
- Chromium and arora doesn't respect disabled states while setting initial
|
|
80 |
choices of select (scatter plot options). Just force proper selection. |
|
339
by Suren A. Chilingaryan
Update of ToDo lists |
81 |
- Opera/Chrome do not resize source popup properly on axes update. If width
|
82 |
of axes tab grows, it just gets out of popup borders. |
|
83 |
- The following case: We made a selection and opened one of the popups. The
|
|
84 |
forced graph update have canceled selection. However, the selected values |
|
85 |
are still present window configuration dialog. Should we cancel them? |
|
86 |
Or should we re-draw selection after update? |
|
87 |
- At the moment on the browser window resize the graphic is not autoupdated,
|
|
88 |
we probably should force update. |
|
286
by Suren A. Chilingaryan
Update of setups |
89 |
|
339
by Suren A. Chilingaryan
Update of ToDo lists |
90 |
Graph
|
91 |
=====
|
|
415
by Suren A. Chilingaryan
Synchronize ToDo lists with zweb |
92 |
- With small graph area (and big left part), the time labels are starting
|
93 |
to overlap at some point (Andreas) |
|
94 |
- After zooming by mouse, when changing to the secondary window or info tab,
|
|
95 |
the "Invalid WINDWO + is specified" returne (Andreas) |
|
339
by Suren A. Chilingaryan
Update of ToDo lists |
96 |
- Prevent complete shadowing (filling) of screen with a single chart in the
|
97 |
multiaxis mode. Try to find such axes and fix the problem by stacking them |
|
98 |
vertically and increasing axis range. |
|
99 |
- Logarithmic axes has problems with drawing 0 values, - where should we
|
|
100 |
start? |